On Thu, December 29, 2005 4:10 am, J. McRee Elrod wrote:
> With the posting of the RDA draft at:
> http://www.collectionscanada.ca/jsc/rdadraftpt1.html
> one can't help wondering reading it just what the impact upon MARC
> development might be.

I'm on record already here as urging caution -- by all means, though,
let's discuss possibilities; IMO actual changes are better left until we
have the whole code.

> In terms of the goal of being written in understandable language, the
> authors and editors have done a [superb] job.

Agreed.

> When I say I prefer ISBD order
> to the new RDA order, what I really mean is that, as I said above, I
> prefer rules in the order in which the elements are encountered on an
> electronic worksheet.  I'm not convinced there is good reason to
> depart from ISBD order (which MARC21 order pretty well follows from
> 245-5XX).  In particular, I find having traced titles and notes mixed
> in with descriptive elements difficult.  There will be a lot of
> scrolling up and down (if one follows RDA), or flipping back and forth
> (if one uses print RDA following a MARC21 worksheet).

I, on the other hand, like the notion of treating each cluster of related
data elements in a set, including notes and/or additional access or
relationships that are related to them.

> One of the basic objectives of the ISBDs is that an item may be
> catalogued once in the country of publication, and then that
> description used internationally.  The RDA substitution of English
> phrases, e.g., "[publisher unknown]" or "[Publisher unknown]" (this
> phrase is inconsistently given), with the phrases in other languages
> introduced in other language libraries, strikes at the heart of that
> basic purpose.

I agree that this is a difficulty; but so it is with notes in records from
non-English cataloguing (e.g. French Canadian materials).  Maybe we have
to work on mechanisms and translation engines to deal with this.  It's
pretty plain to me that cataloguers without a reasonable working knowledge
of the language of their audience can't be expected to create records in
another language.

> What language would we use for WHO, a library in a
> bilingual country serving an international patronage?  Latin
> abbreviations are a useful compromise in a multilingual situation.

But, much as I regret it, in other environments that's not, or no longer, so.
>
> Bilingual catalogues use 040$b to determine the language used for
> display constants, e.g., a record with 040$bfre would have French
> phrases used for the 246 2nd indicators.  This allows us to have one
> record serving libraries with English, French, or bilingual
> catalogues.

What about notes?
>
> There is also the matter of the relative length of "et al". vs. "and
> others", "s.n." vs. "publisher unknown", in terms of what appears in a
> one line display or printed new titles list.

I think the day of abbreviations, apart from the most obvious (maybe
"etc." is the only one that qualifies) has passed -- and I include
geographical abbreviations.

> In most cases experienced cataloguers will know what MARC21 tag
> applies to which element of RDA description.  But sometimes it is not
> so clear.  For example for, 1.2.3 multilevel description, in which the
> description of the whole and the description of the parts are in a
> single record, even with the example at D.1.4, it is not clear to me
> what elements will be coded how.

I agree we need help here: I too would like to know how multilevel records
will be coded -- perhaps as a set of related records, with distinct
subject and added entry access for each component?  Can anyone sketch how
this might work?

> The distinction between major title changes (1.3) and minor (as listed
> in 2.3.1.12 b) does not include as minor a single issue change as
> recently discussed on Autocat, and covered by CONSER but not AACR2.
> Perhaps this deserves inclusion, as well as allowing the reopening of
> a record when a serial reverts to an earlier title, as opposed to
> having two records with the same 245 for the same publication as now.
> We've never had a customer which will accept the three records for
> Atlantic/Atlantic monthy for example, another minor change which should
> be added, with perhaps 247 allowed for more than integrating
> resources.

I would argue that for most instances where a title changes then reverts
(or an issuing body changes and reverts) data should be consolidated into
a single record -- on the principle of user convenience.
>
> The list of mandatory elements (1.4) includes Extent 300 (even though
> not covered in this draft), but does not include place of publication
> 260$a .  The latter is a major oversight from my point of view.  Our
> law firm library customers consider jurisdiction of publication far
> more important than name of publisher - some of whom are
> international.

This perpetuates what I believe is a deficiency in the AACR2 Level One
specification.  I would never argue for omission of publisher from even
the briefest record; but to omit place is to run counter to user
expectation.  Place is always one of the requirements in every guide for
writing citations, for lists of references and so on.  The great older
catalogues such as the British Museum gave place but, often, not
publisher.

> It's good to see in 1.6.1.1 such capitalizations as "e-Commerce", and
> "www. ..." allowed at the beginning of title transcription.

To me, the odd thing is that recording of such distinctive usages was
thought to require permission.

> The rule of three is now optional (2.3) for statements of
> responsibility.  Presumably if the option is not adopted, the number
> of 700's will greatly increase, as well as the length of 245$c.

Maybe there should be a trial to see how much this is so, and how it
affects different disciplines; perhaps also to test the effects of
skipping authority creation where there is no conflict and no references
or other authority information are called for?  Authority work based on a
single occurrence may be a luxury we can no longer afford.  There is a
MARC 720 tag for uncontrolled names.  For names of this type, I wonder how
often authorities would have to be created later, and how often changes
would be necessary?

> Monographs also have earlier and later
> titles in successive editions, but these relationships are not
> coverend in this section, so one assumes 247 and 780/785 will not be
> extended to monographs.  The 780/785 way of relating earlier and later
> editions would seem to me a simple way of meeting FRBR objectives.

Why not?
>
> In 2.8.0.3, as mentioned earlier, jurisdiction for place of
> publication is to be transcribed if present (in contrast to the rarely
> observed AACR2 provision), but lacking jurisdictions are not supplied.
> This is something we have to do for our international customer base.
> We can't expect an European or Asian patron to know the jurisdiction
> of North American cities.  Also, if we have "London, Ont." it seems to
> me we should have "London [England]".

Perhaps direct transcription isn't the right approach here; rather, to
render the place (and the publisher's name), based on what's present in
the item.  Maybe jurisdiction can better be supplied automatically based
on the country of publication data recorded in the fixed field coding
(which I would like to have on the editing screen adjacent to the
bibliographic data field to which it applies).
>
> Some changes in 2.9.1.3 for date of publication seem wise: "[1800s?]"
> seems clearer than the present "[18--?]".

Even better might be "[19th century]".

> But why not "[late 1800s or
> early 1900s]" as opposed to "[date unknown]", if that is the case?
> Certainly the cataloguer with item in hand is better able to determine
> that than the patron at the catalogue.

An approximate date is far better than none.
>
> Unfortunately 2.3.1.1 still defines alternate title as being part of
> title proper.  If so, why does it begin with an upper case letter?
> That first portion of the title should be followed by a 245$hGMD, and
> the alternate title should be coded 245$b, as happens with all the
> other titles other than first title proper, whether parallel title, or
> a later title in an item without a collective title (the latter being
> a rule change).

In my view, every separately identified title element ought to be
separately subfielded.  I think the ISBD rule for alternative titles is
wrong -- and often gives rise to errors in formulating uniform titles,
which (by LCRI) should exclude alternative titles; and (especially in
other languages) mistakes in cataloguing do occur here.

> I can see a lot of annotating of the printed version
> with MARC21 field tag numbers.  Wouldn't an appendix doing that be
> helpful?

Or better, since there will be consequent MARC changes, a separate
explanatory document?

Hal Cain, still with much reading of the RDA Part I draft to do
Joint Theological Library
Parkville, Victoria, Australia
[log in to unmask]